CHAPTER 3 METHODOLOGY
3.6. Instruments
Two self-report questionnaires were used in this study: (1) a questionnaire pertaining to personal attributes that previous research indicated may be relevant to the phenomena of interest, and (2) the Professional Quality of Life Scale – Version 5 (ProQOL) (Stamm, 2009).
3.6.1. The personal attributes questionnaire
This questionnaire was designed to investigate selected personal attributes found in literature to influence positive and negative sequelae of caring work. A database of factors listed in existing research was created (see Appendix 2), and from this a list of areas to be covered in the questionnaire was selected. The resulting questionnaire was
also translated into Zulu, as it was anticipated that this would be the first language of many participants. The Zulu form of the questionnaire was checked by a second translator to ensure accuracy. Both translators were first language Zulu speakers from the same region of KwaZulu-Natal as the research participants, in order to avoid potential problems with regional language differences. A copy of the final questionnaire in both forms can be found in Appendix 4.
3.6.1.1 Description
Specific areas of enquiry covered in this questionnaire were as follows:
(i) Demographics section (a) Gender
(b) Age
(c) Current marital status – whether married, in a long-term relationship, divorced, separated, widowed, single/ never married
(d) Scholastic education – highest grade reached in school
(e) Post school qualifications – any certificate courses, diplomas or degrees obtained after leaving school
(f) Occupational status – whether currently employed, studying, retired, voluntary workers or unemployed
(g) Personal history of HIV and AIDS – whether caregivers or somebody close to them has been directly affected by HIV or AIDS
(ii) Attributes related to work as a caregiver
(a) Experience – how long the respondent has worked as a caregiver
(b) Remuneration status – whether voluntary/unpaid, given stipend, or employed as caregiver
(c) Caseload – estimated number of clients seen per week
(d) Level of training – the length of training the caregiver has received
(e) Perceived adequacy of training – whether caregivers feel that their training prepared them adequately for the work they are doing
(f) Ongoing training and development – whether caregivers have access to such (g) Access to therapist or counsellor if required by caregiver
(h) Personal identification with clients – the degree to which caregivers feel that they are similar to their clients
(i) Role clarity – the degree to which they are certain of the extent and limits of their role
(j) Sense of personal efficacy – whether respondents perceive themselves to be performing their caregiving work well
(k) Sufficiency of referral networks – extent to which caregivers are confident in their ability to refer clients to appropriate sources of assistance in cases where they have needs that fall outside the limits of the caregiver‟s role
(l) Secondary stigma – degree to which caregivers perceive that their work exposes them to secondary stigmatisation.
(m) Social support – whether caregivers perceive those close to them as a source of support and strength
(n) Supervision and mentoring – whether they receive help and advice regarding clients from people within their organisation
(o) Self-care – extent to which caregivers consciously engage in activities that are pleasurable and relaxing.
3.6.2 Professional Quality of Life Scale – Version 5 (ProQOL) 3.6.2.1 Description
The ProQOL 5 consists of 30 items measured in a 5-point Likert scale self-report questionnaire. Results are scored on three subscales, namely Secondary Traumatic Stress, Burnout and Compassion Satisfaction (see Table 3.1 below). This measure is the current version of the Compassion Fatigue Self Test developed by Figley in the 1980s (Stamm, 2009). It is the most widely used measure used to assess the positive and negative effects of working with people who have experienced trauma (Stamm, 2009). The ProQOL was developed in English, but has previously been translated into eight additional languages (Stamm, 2009).
For this research, the ProQOL was also made available in Zulu, as it was anticipated that this would be the first language of many of the participants. To ensure an accurate and accessible Zulu version while also maintaining the integrity of the measure, a procedure of translation and back-translation by independent translators was followed.
Where it proved inappropriate or impossible to translate particular items on the
measure directly, the translation was done in consultation with the researcher so as to reflect the intended meaning of the items as accurately as possible.
Table 3.1 Definition of Subscales of Professional Quality of Life Scale ProQOL 5 Subscale Items Definition
Compassion Satisfaction (CS)
10 Positive sequelae of helping others; i.e. pleasure derived from doing work well (Stamm, 2009).
Burnout (BO) 10 Negative sequelae of helping others, specifically associated with a sense of futility and helplessness (Stamm, 2009).
Secondary Traumatic Stress (STS)
10 Negative sequelae of helping others, specifically associated with fear and work-related trauma (Stamm, 2009).
3.6.2.2 Validity and reliability:
Construct validity is reported as good, with over 200 published research papers and more than 100 000 internet articles to date (Stamm, 2009). The ProQOL or earlier versions thereof have been used in many of these publications (Stamm, 2009).
Although it is reported that there is shared variance between the Burnout (BO) and Secondary Traumatic Stress (STS) subscales, the measure‟s author contends that this can be explained as a reflection of the distress that is common to both these conditions. The two subscales measure separate constructs and are distinct in that the STS scale addresses fear while the BO scale does not (Stamm, 2009).
The concise ProQOL manual reports alpha reliabilities and standard errors as detailed in Table 3.2 below.
Table3.2 ProQOL Subscale Alpha Reliabilities and Standard Errors (Stamm, 2009)
Subscale Alpha reliability Standard error
Compassion Satisfaction α = .88 (n = 1130) .22
Burnout α = .75 (n = 976) .21
Secondary Traumatic Stress α = .81 (n = 1135) .20
The measure‟s item-to-scale properties are reported to be good, as no single item either adds to or subtracts from the overall quality of the scale (Stamm, 2009).
3.6.2.3 Normed population
Although ProQOL normative data does not exist for the specific population of interest in this study, the norms contained in the manual reflect a large database created from multiple studies in international settings. No scale score differences were observed between different countries (Stamm, 2009). According to Stamm (2009), the ProQOL is appropriate for use in any population where there is a possibility of being exposed to another person‟s potentially traumatic material as a part of either paid or volunteer work.
3.6.2.4 Scoring and interpretation
First, scoring for 5 items on the Burnout (BO) subscale is reversed (see appendix).
Then the ratings for items on each subscale are summed and converted to T-scores.
The mean score for all subscales is therefore 50 with a standard deviation of 10. The test manual advises that scores are best regarded in their continuous forms, as the use of cut-point scoring tends to lead to the commission of type I errors; however cut scores are provided in the manual for screening purposes. A summary of subscale score interpretation is provided in Table 3.3 below.
Table 3.3 Interpretation of ProQOL Scores
Subscale Interpretation of scores Compassion
Satisfaction (CS)
High scores reflect high levels of satisfaction with helping work, characterised by enjoyment and feeling invigorated, effective and successful.
Burnout (BO) High scores reflect high risk of burnout, characterised by feelings of futility, ineffectiveness, exhaustion and being overwhelmed.
Secondary Traumatic Stress (STS)
High scores reflect high risk of secondary traumatic stress, characterised by feeling fearful, experiencing symptoms of trauma, and being preoccupied with thoughts of people one has helped.
(Stamm, 2009)